Refactor IntegerMulAction

[jdemeyer]

Refactor IntegerMulAction in preparation for #24247:

1. Add an abstract base class IntegerAction which will also be used to implement IntegerPowAction in Implement __pow__ in the coercion model #24247.

2. Add a new helper function parent_is_integers to check if some type/parent represents the integers.

3. Instead of the hacks in discover_action and verify_action involving an explicit IntegerMulAction check, return an IntegerMulAction for anything satisfying parent_is_integers().

Component: coercion

Author: Jeroen Demeyer

Branch/Commit: a359862

Reviewer: Travis Scrimshaw

Issue created by migration from https://trac.sagemath.org/ticket/24467

[jdemeyer]

Branch: u/jdemeyer/ticket/24467

[jdemeyer]

Commit: 85cabe7

[jdemeyer]

New commits:

85cabe7

Refactor IntegerMulAction

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

a2d60ef

Also verify IntegerAction instances

[sagetrac-git]

Changed commit from 85cabe7 to a2d60ef

[tscrim]

Reviewer: Travis Scrimshaw

[tscrim]

comment:5

The · is not ascii nor latex: n · a = a + a + ... + a. So if you change that to \cdot (personally, I would also change the ... to \cdots), then you can set a positive review on my behalf.

[jdemeyer]

comment:6

Replying to @tscrim:

The · is not ascii nor latex

Is that a problem really? The documentation builds fine and I think that · is more readable than \cdot in text mode.

[sagetrac-git]

Changed commit from a2d60ef to a359862

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

a359862

Use \cdot instead of ·

[tscrim]

comment:9

Replying to @jdemeyer:

Replying to @tscrim:

The · is not ascii nor latex

Is that a problem really? The documentation builds fine and I think that · is more readable than \cdot in text mode.

With the non-ascii characters, the pdf docs tend to have difficulty if the encoding is not declared (in particular, the non-ascii hyphen comes to mind). What we probably should do is add a bunch of the unicode versions of certain latex commands to the console doc replacer since we generally assume terminals support unicode.

Thanks for the change.

[jdemeyer]

comment:10

Replying to @tscrim:

What we probably should do is add a bunch of the unicode versions of certain latex commands to the console doc replacer since we generally assume terminals support unicode.

No, we do the opposite: replace certain Unicode characters by latex when building the PDF docs. In src/doc/common/conf.py you will find for example

\DeclareUnicodeCharacter{00B7}{\cdot}

saying that · should be replaced by \cdot. So · in docstrings should work just fine.

[tscrim]

comment:11

I see. I wasn't aware of that. Thank you.

[vbraun]

Changed branch from u/jdemeyer/ticket/24467 to a359862